> ## Documentation Index > Fetch the complete documentation index at: https://docs.purna.ai/llms.txt > Use this file to discover all available pages before exploring further. # Sequence Alignment Component > Inspect biological sequences and prepared multiple sequence alignments, select regions, and understand supported files and preview limits. The Sequence Alignment Component lets you inspect nucleotide and protein sequences inside your Purna workspace. Open a supported file from **Files** or a chat attachment to explore it alongside your conversation. A multiple sequence alignment (MSA) places related sequences in rows, with corresponding positions in columns. Gaps preserve the alignment supplied in your file. The viewer displays that alignment; it does not decide which sequences should be aligned or calculate a new alignment. ![Sequence alignment displayed in Purna's workspace for inspecting aligned residues and regions](https://cms.purna.ai/_emdash/api/media/file/01M6GZH8095HHT47RVC7PV7BZ5.01M6GZH9M46307Z3V4Q4JY4E0Z.webp) Explore a prepared sequence alignment inside your workspace. ## What you can do | Capability | What you can inspect or change | | - | - | | Alignment view | Compare the aligned residues and gaps already present in your file | | Sequence view | Read individual records, descriptions, residue counts, and basic sequence statistics | | Navigation | Scroll through rows and columns, zoom horizontally, fit the alignment, and find a sequence by identifier | | Colors | Choose nucleotide, Clustal, protein Taylor, or no residue coloring | | Conservation | Enable the optional conservation display to inspect agreement across alignment columns | | Selections | Highlight a rectangular block of alignment cells or a column range across all rows | | Workspace controls | Resize the pane, expand the viewer, switch between open tabs, and download the original file | This component provides sequence and alignment inspection. It does not provide the full range of tools found in a dedicated alignment editor or genome browser. ## Supported files | Format | Supported extensions | Requirements and opening behavior | | - | - | - | | FASTA | `.fa`, `.fasta`, `.faa`, `.fna`, `.ffn`, `.frn`, `.fas`, `.mfa` | Single records or records with different lengths open in **Sequence** view. Equal-length records containing gaps open in **Alignment** view. Equal-length gapless records open in **Sequence** view with an optional **Alignment** toggle. | | Explicit aligned FASTA | `.afa`, `.afasta`, `.aln-fasta` | At least two records with equal alignment lengths; opens in **Alignment** view. | | Clustal | `.aln`, `.clustal`, `.clw` | One complete alignment with a valid Clustal-style header. Interleaved blocks must contain the same identifiers and equal-width fragments. | | Stockholm | `.sto`, `.stk`, `.stockholm` | One complete alignment beginning with `# STOCKHOLM 1.0` and ending with `//`. Interleaved blocks are supported. | | A3M | `.a3m` | At least two records with the same number of match columns. Lowercase insertion residues are expanded into aligned display columns. | Extensions are case-insensitive. Use an extension that matches the file's actual contents. For example, a FASTA-formatted alignment saved with an `.aln` extension needs the appropriate FASTA extension; changing a filename does not perform a format conversion. ### Prepare your alignment * Supply an alignment you have already calculated when you want to compare corresponding positions. Equal sequence lengths alone do not establish a valid biological alignment. * Give every sequence a unique identifier. In FASTA and A3M headers, the first word is the identifier; the remaining text is its description. Descriptions alone do not distinguish duplicate identifiers. * Keep all rows, gaps, and blocks intact. Upload one complete alignment per file. * Use UTF-8 text and valid sequence characters for the selected format. Numeric quality scores, spreadsheet exports, or binary data are not sequence rows. * For A3M, keep uppercase match residues, deletion dashes, and lowercase insertions in their original roles. Do not uppercase the file to make its rows appear equal in length. * If your A3M export contains additional non-sequence records, such as numeric confidence annotations, export a sequence-only viewing copy before uploading it. When exporting from another tool, choose a supported format explicitly. A ZIP archive, a renamed FASTQ file, or a table of sequences is not a prepared alignment. ## How your file is prepared for viewing When you open a supported file, Purna checks that it is complete, readable, and within the preview limits. It checks the sequence identifiers, row structure, and format-specific alignment requirements before displaying it. Ordinary FASTA can use **Sequence** view when it is not an alignment. Purna does not add gaps or align unequal-length records automatically. Files explicitly marked as alignments must satisfy their format's alignment requirements. For A3M, insertion positions are expanded for display. The displayed grid can therefore have more columns than an individual input sequence. Your original file is unchanged, including its identifiers, residues, letter case, and gaps. You do not need a companion index, a second upload, or a manually prepared preview file. Opening the file does not trim sequences, infer homology, calculate a tree, perform variant calling, or start a biological analysis. To create or convert an alignment, request that work separately and review the output before opening it. ## Working with your data ### Open and arrange the viewer 1. Upload your supported file through **Files**, or attach it to a conversation. 2. Select the file and wait for **Opening sequence file…** to finish. 3. Use **Sequence** or **Alignment** when the corresponding view is available. 4. Resize the workspace pane using its divider, or use the expand icon for more room. Use **Download** in the file toolbar to retrieve the original. If fullscreen is unavailable, resize the notebook pane instead. ### Inspect an alignment * Scroll to move through sequence rows and alignment columns. * Use **−** and **+** to change the horizontal zoom, or **Fit** to bring the alignment into view. * Use **Find sequence…** and **Find** to jump to the first matching sequence identifier. Search is case-insensitive and accepts part of an identifier; it is not a residue or motif search. * Choose a **Colors** scheme to distinguish residues. * Enable **Conservation** when you want to inspect agreement across columns. Leave it off when you do not need it, especially for larger alignments. * Use **Reset** to return zoom and position to their defaults, restore default coloring, and clear the current selection. Colors and conservation help you inspect an alignment. They do not establish functional importance, evolutionary relationships, pathogenicity, or the quality of the alignment. ### Select cells and regions An alignment **cell** is one row-by-column position containing a residue or gap. It is not a biological cell, as in the Spatial Omics Viewer. * Hold **Shift** and drag across the alignment to select a rectangular block spanning particular sequences and columns. * Drag across the position ruler to select a column range across all rows. * Read the selected column range and row information shown above the alignment. * Select **Clear selection** to remove the highlight. Resetting the view also clears it. Selections identify positions in the displayed alignment, including gaps and expanded A3M insertion columns. An alignment column is not necessarily the same as a residue number in an ungapped sequence or a coordinate in a reference genome. Record the sequence identifiers, coordinate convention, and relevant reference when requesting analysis of a region. Selections are temporary highlights. They do not edit, crop, export, save, or automatically send the selected data to an agent. ### Read individual sequences In **Sequence** view, use **Previous**, **Next**, or **Find record…** to navigate records. Record search checks identifiers and descriptions. The selected record displays its description, position-numbered sequence text, residue count, and inferred nucleotide-like or protein-like coloring. Residue counts exclude gaps and stop markers. For nucleotide-like data, GC percentage uses unambiguous A, C, G, T, and U bases as its denominator; ambiguity symbols and gaps are excluded. Alphabet inference can be ambiguous, so verify the sequence type and source before using these values in an analysis. ## Work with an agent Agents can work with your files for analysis, statistics, and scripts. You can ask for tasks such as counting gaps, calculating conservation statistics, checking identifiers, converting a supported alignment, or preparing a smaller viewing copy. Review the method and any new output file before using it. Agents cannot view the graphics or slides displayed inside scientific components. They can read and interpret selections when you provide the selection as readable data, sequence identifiers, and a clearly defined region. The Sequence Alignment Component does not automatically attach its graphical selection to chat. When asking about a region, provide the file, the relevant sequence identifiers, the column range, whether positions include gaps, and the question you want answered. Ask the agent to confirm the coordinate convention before running calculations. A request such as "analyze what I selected" alone does not give the agent that context. > [!WARNING] > > Purna is designed for a human in the loop. Do not rely completely on the agent. Inspect the alignment yourself and review the agent's assumptions, coordinate mapping, statistics, scripts, and conclusions. An agent's answer does not replace scientific judgment or experimental validation. ## Size and performance limits | Preview limit | Maximum | | - | - | | Input file | 10 MiB | | Sequence records | 10,000 | | Alignment columns | 100,000 | | Displayed alignment positions | 5,000,000, calculated as rows × columns | All applicable limits must be met. For example, 5,000 sequences × 1,000 columns reaches the displayed-position limit, even though both dimensions are below their individual limits. A3M limits apply after insertion columns are expanded. Ordinary FASTA within the file and record limits can still open in **Sequence** view when its alignment would exceed the column or displayed-position limit. Explicit alignment formats that exceed those limits do not open as an alignment. A successful upload does not guarantee a preview. The complete file must load and pass validation before the alignment appears; incomplete downloads and partial alignments are not progressively displayed. Large files or wide alignments can take longer to open and consume more memory, even within these limits. For a slow or oversized file, prepare a smaller viewing copy using selected sequences or a clearly defined region. Preserve the original and document any filtering or trimming. Repeatedly reopening an unsupported file does not convert it or remove the limit. ## Temporary state Your sequence file remains unchanged. Colors, selected records, zoom, position, and highlighted regions are temporary. Switching between open tabs preserves the temporary view for the same file. Closing the tab or reloading the page discards it. Use **Reload sequence file** to reread the source when needed; if the underlying file has been replaced, review its contents and make a new selection. This component does not save annotations, personal labels, edited alignments, or selection history. Downloading retrieves the original file, not the highlighted region or a screenshot. ## Files and workflows that will not work * Compressed sequence files such as `.fa.gz`, `.aln.gz`, and `.a3m.gz`, or alignments packed inside ZIP or TAR archives. Decompress and upload the supported text file. * FASTQ, GenBank, EMBL, SnapGene, and chromatogram files. These need different viewers or an explicit conversion suited to your task. * PHYLIP, NEXUS, MSF, PIR, A2M, EMF, and other formats outside the supported list. Export a supported alignment instead. * CSV, TSV, JSON, and ordinary text files renamed to an alignment extension. CSV files keep their table preview. * Multiple concatenated alignments in a single file. Split them into one complete alignment per file. * Duplicate identifiers, missing sequence rows, inconsistent alignment blocks, malformed text, or truncated files. * Unequal match-column counts in A3M, or unequal row lengths in explicit aligned FASTA. * Tree files, domain tracks, genome annotations, or quality-score records supplied as alignment files. This component does not provide tree construction, annotated genome browsing, sequence editing, or quality-score analysis. ## Troubleshooting | What you see | What to check | | - | - | | **Alignment** is unavailable | Ordinary FASTA needs at least two equal-length rows within the alignment limits. Single or unequal-length records remain viewable in **Sequence** mode. | | A gapless FASTA opens in **Sequence** | This is the default. Choose **Alignment** if you know the equal-length records represent an existing alignment. | | An expected header or unsupported-format error | Match the extension to the actual file format. Export or convert the file explicitly; do not rename unrelated data. | | Duplicate identifiers | Check the first word in every FASTA/A3M header. Assign unique identifiers in a separate viewing copy while preserving their mapping to the original records. | | Inconsistent rows or incomplete blocks | Export one complete alignment. Check that interleaved blocks contain the same identifiers and equal-width fragments, and that Stockholm ends with `//`. | | An A3M match-column or invalid-character error | Check match versus insertion case, deletion dashes, and any non-sequence annotation records. Export a compatible sequence-only A3M copy. | | A size or displayed-position limit error | Check file bytes, sequence count, and expanded alignment width. Use a smaller viewing copy; a short A3M file can expand into a much wider grid. | | Loading fails, parsing takes too long, or a download is incomplete | Wait for the upload to finish, check your connection, and use **Reload sequence file** or **Retry** when offered. Try a smaller complete file if the problem persists. | | Access denied, file unavailable, or file changed | Confirm access to the file and project. Reopen the current file from **Files** and recheck the selection if the source was replaced. | | The renderer cannot start or shows a blank panel | Retry or reload the page after noting temporary selections. Try a smaller alignment or a current desktop browser. | | Scrolling or interaction is slow | Turn off **Conservation**, test fewer sequences or columns, and close other large previews. Files within the limits can still be demanding. | | **Find** does not locate a motif | Alignment search finds sequence identifiers; record search also checks descriptions. Ask for a separate sequence-pattern analysis when needed. | | Selection coordinates do not match a sequence or reference | Distinguish displayed alignment columns, ungapped residue numbers, and reference coordinates. Include A3M expansion and gaps when defining the mapping. | | A selection disappears or an agent does not know it | Selections are temporary and are not automatically shared. Provide the file, identifiers, region, and coordinate convention in chat. | Return to [Scientific Components](https://docs.purna.ai/scientific-components/overview.md), or see [Files](https://docs.purna.ai/guides/files.md) for uploading and organizing your data. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.